Webinar

Fireside Chat: FinCrime Agent Benchmark and Evaluating AI for AML

What does it take for AI to perform reliably in AML? The FinCrime Agent Benchmark (FCA Bench) shows how results change when models work on their own versus within an agentic harness built for AML/KYC. In this session, compliance practitioners and AI implementation experts discuss what the findings mean for institutions evaluating AI solutions or building their own.

  • What FCA Bench reveals about AI’s strengths and limitations in AML/KYC workflows

  • How an AI harness affects decision accuracy and consistency

  • How the quality, completeness and freshness of risk data affect an agent’s ability to identify and assess risk

  • What goes into building an effective agent beyond choosing a model, including data, policy instructions, workflow controls, testing and governance

  • How to evaluate an AI solution before deployment, and what institutions need to consider when deciding whether to build or buy

Meet the Speakers

Will is Sr. Director of CX-Global Financial Crimes Operations at Coinbase. He brings 20+ years in fraud and AML risk across banking and law enforcement and is a former co-chair of the ACAMS Northern California and Greater Phoenix Chapters.

Peter Piatetsky

Co-Founder & CEO

Castellum.AI

Jared Terry

Global Head of Financial Crimes Program Integrity and Effectiveness

Stripe

Will Voorhees

Sr. Director CX-Global Financial Crimes Operations

Coinbase

Peter leads strategy, growth and product at Castellum.AI, working closely with clients to implement risk-aligned solutions. Prior to co-founding Castellum.AI, Peter served at the US Treasury Department.

Jared leads the Financial Crimes Program Integrity and Effectiveness function at Stripe, where his team sets the investigative standards, quality frameworks, and AI-enabled tooling strategy that enable Stripe's program to scale across sanctions and AML globally. He brings 15+ years of experience.

Featured Resource
See the Full Benchmark

FinCrime Agent Benchmark

The first benchmark to evaluate how AI models apply an institution’s AML/KYC policies, identify risk and adjudicate alerts. What’s inside the benchmark:

  • Model leaderboard: How 9 leading LLMs perform across 500 sanctions, PEPs and adverse media cases.

  • Standalone vs. harnessed results: How each model performs on its own and through Castellum.AI’s agentic harness.

  • Screening effectiveness: How many expected hits each model finds before adjudication begins.

  • Decision consistency: Whether a model reaches the same decision when it runs the same case 3 times.